Experimental Digital Twin Framework for Cybersecurity Risk Analysis

Digital Twins (DTs) are increasingly adopted across domains such as smart cities, Industry 4.0, connected agriculture, and education, where they enable real-time monitoring and decision-making. However, securing end-to-end DT architectures (panning IoT data acquisition, data governance, and AI-driven analytics) remains a significant challenge. In particular, the lack of experimental and reproducible environments hinders systematic cybersecurity evaluation and research.

This paper proposes a baseline experimental digital twin environment dedicated to cybersecurity analysis, with a focus on machine learning–based anomaly detection. We design and implement a minimal yet realistic DT pipeline that ingests sensor data, enforces data governance constraints, and integrates lightweight AI components for monitoring and threat detection. Representative cyber threats are simulated to expose security issues that are specific to DT architectures.

The main contribution of this work lies in providing a reproducible, pedagogical experimental setup that serves as a foundation for further research. The proposed baseline enables readers to understand core DT cybersecurity challenges, replicate the experiments, and extend the environment toward sector-specific digital twins and more advanced machine learning–based security mechanisms.

Coruña, Spain, November 23-25, 2026